10 M. Melloni on the Radiations of Incandescent Bodies, 2 
j “instead of being absolutely pure, or even nearly so, the gradations 
of as image formed by the central compartment contain citer 
colo 
I had already proved, in 1843, that the colors obtained by ¢ or- 
dinary prisms at a distance less than two metres, are composed of 
a mixture of colors belonging to the spectra of the different ele- 
mentary bands of the prism; and that the red, the violet, and_ 
consequently all the other prismatic colors of the two extreme 
elements, are so much the nearer to the centre as the observation 
is made nearer to the prism. Now, in the experiment of Sir 
Brewster the prism is very close to the eye,—the spectrum obs ; 
tained in this experiment must necessarily be formed of very um- ~ 
pure tints; and the color which appears in a given zone, which 
has lost by absorption the dominant tint, does not belong toa 
ray of the same refrangibility as the tint absorbed, but rather to 
colors of the elementary hacia of the superior or inferior —_ 
of the prism 7 
To prove this directly, I have repeated the fundamental exper- 
iment of Sir D. Brewster. It consists, as is well known, in in- 
terposing between the eye and the spectrum, produced by the re- 
fracted image of a luminous object seen through the prism, a slip 
of glass deeply colored blue by the oxyd of cobalt. The sp 
trum was formed from the light of a circular aperture, ten milli+ 
metres in diameter, made in a metallic plate placed in the shutter 
of a dark chamber. The prism was of flint glass, equilateral, 
twenty-five millimetres wide, and sufficiently pure to produce 
distinctly the black lines of Sannin It was attached to its 
support by one end, placed fifteen feet from the window, and fix- 
ed horizontally in the position of minimum deviation. | Its ante- 
rior face was covered for one-third of its extent with India ink. 
From the middle of this there was removed a longitudinal space 
which ran from end to end of the blackened band, leaving free 
and uncovered a horizontal line little more than a millimetre in 
width. The slip of blue glass covered only went e of the 
prism counting from the painted extremity. te 
Things being thus arranged, I observed the image af. the aper- 
ture successively through the uncovered part of the prism, and 
then through the two portions on which the blue glass was pla- 
ced. ‘I'he first operation gave me the normal spectrum; the se- 
cond, made after the method of Brewster, furnished a complex 
spectrum ; the third, a spectrum arising from a little portion whieh _ 
may be regarded as the mean longitudinal element of the prism: 
a on menpening the first image with the second, I observed. 
ena of por gaet and “obec zones s0:rek described — 
One 
eating then the third: ne with the se~ 
ein Oe fe vk 
Se ed aes 
Fp eee ae 
